Our Take
MikeOSS exposes how little competitive advantage legal AI vendors actually have in their software features alone.
Why it matters
Law firms and corporations can now consider building legal AI tools internally instead of paying vendor subscriptions, forcing established players to compete on implementation and support rather than core functionality.
Do this week
Legal tech buyers: audit your current vendor contracts before renewal to understand what you're paying for beyond base features.
Developer copies major legal AI platforms
Will Chen released MikeOSS on GitHub last week, an open-source project that claims feature-equivalence to Harvey and Legora (per LawSites reporting). The project includes an AI assistant, tabular review functions, and reusable workflows. The development community response has been described as viral, though no specific adoption metrics were provided.
The release highlights how AI-assisted coding tools like Claude Code have reduced development barriers. According to industry observer Ken Crutchfield, "the cost of developing feature equivalent software has just decreased by one or two orders of magnitude."
Build versus buy equation shifts toward build
MikeOSS represents a broader trend where open-source alternatives emerge for previously proprietary legal technology. This mirrors patterns in other software categories where Google open-sourced Android and Meta released Facebook's user-experience technology.
The legal industry has historically struggled with knowledge sharing due to structural incentives. Origination credit, client portability concerns, and confidentiality requirements create motivations that run counter to collaboration. Partners often protect their "secret sauce" to maintain competitive edges, even when firm-wide sharing would benefit the organization.
However, large language models have been systematically collecting legal knowledge, creating what Crutchfield calls "secret sauces that rival any individual attorney's proprietary know-how." As AI absorbs and redistributes legal expertise, the competitive advantage shifts from holding knowledge to operationalizing it effectively.
Implementation costs remain the real barrier
Building in-house legal AI involves more than replicating features. Organizations need implementation, integration, training, and ongoing support capabilities. Security requirements include SOC 2 certifications, clear release processes, testing protocols, and separation of duties.
Some corporations restrict which open-source technologies can access their data, adding compliance overhead. The political costs of deploying homegrown systems often exceed initial estimates, making vendor purchases safer organizationally.
The real disruption isn't that open-source tools can match Harvey or Legora's features. It's the exposure of how little competitive advantage actually comes from software functionality versus execution. As Crutchfield notes, winners won't be firms with the best recipes, but "the ones running the best kitchens" to deliver services securely and repeatably.